Improving the utility of national vegetation and ecosystem maps

Background/Question/Methods

The US National Vegetation Classification (NVC) has helped to standardize and accelerate vegetation and ecosystems mapping efforts around the globe.  These national and continental efforts rely on satellite image processing techniques to best analyze the deep levels of spatial and classification detail within the large data volumes required for continental scales.   These data are important for and needed by day-to-day conservation planners, managers and organizations everywhere.  These individuals and organizations are generally skilled in the use of standard Geographic Information System (GIS) vector and database structures to carry out resources management, conservation, education and advocacy tasks.  Unfortunately, the normal distribution of NVC national datasets is as huge raster catalogs of a size and type compatible with large image processing efforts, but beyond the typical GIS abilities of most planners and conservation organizations.  The online map offerings for NVC data are generally single-access applications containing raster basemaps, which by necessity must be opaque, limiting their use in a GIS or map visualization scenario, where details of surface imagery often need to be combined with thematic overlays in different ways for different questions.

Results/Conclusions

To make detailed national NVC mapping efforts more useable to the average conservation group, they need to be converted from raster basemaps into vector overlays, suitable for use over other basemaps such as high-resolution natural imagery or sequences of historical imagery.   Converting from raster to vector at this scale presents many challenges, not the least of which is assembling a national vegetation vector database of almost a billion features that will be responsive when accessed online, easy to distribute, and created and maintained using simple, readily-available technologies.  Additional challenges include: designing cartography for nearly 1,000 legend classes, refining vectorization processes to maximize fidelity with the landscape, creating a single level of boundaries at the finest scale that can still function at coarser scales to permit a proper spatial representation of the nested taxonomic hierarchy of the NVC,  maintaining source data area measurements in their original projection as online maps are generated in other projections,  enabling offline field use and automating a process to allow rapid updates as additional image processing and classification refinement continues.